This PR improves the dataset card for MULocBench by:
- Adding a link to the paper: https://huggingface.co/papers/2509.25242
- Adding an introductory description to reflect the dataset's purpose as described in the paper abstract.
- Adding `task_categories` (`text-retrieval`) and relevant `tags` (`code`, `software-development`, `issue-localization`, `bug-fixing`).
- Specifying the `language` as English.